Re: large hard disks (again)
From: Kenneth Scharf [EMAIL PROTECTED] . However this is NOT the physical geo, which IS reported by the bios and stamped on the drive. The geometry printed on the label isn't always the actual IDE geometry; it's typically limited to 16383 cylinders (for BIOSes that can't handle the full size). Daniel
large hard disks (again)
I have posted this before, how to fdisk a 8.4gb drive. Problem was that with slink fdisk saw only 1024 cy, 255h, 63s for a total of 8.4gb. If I booted a RedHat 6.0 or Mandrake 6.0 CD (also stormix) their partition utilities (fdisk, disk druid, etc) would report the correct number of cyl's (2100) to give a 17.2gb total disk size. After partitioning with Stormix I stopped the install, and booted my slink cd and then installed slink (skipping the partition step, which was already done). This worked. Now if I run fdisk, it still reports the wrong(?) number of cyl's but does show the correct partition sizes (p command). It also complains about different logical and physical starting and/or ending cyl numbers for the last two partitions on the disk. However if I go to expert mode (x) and change the number of cylinders (c) to 2100, the p command now gives the exact same printout as fdisk under Mandrake or Redhat fdisk. So...why does fdisk get the correct size info from the kernel under Mandrake or Redhat? Can it be a difference between the 2.0.36 kernel and the 2.2 series? Or is it a later version of fdisk itself (BTW cfdisk and sfdisk behave the same way...) I now think the difference is in the 2.2 kernel. Any ideas? I now know I can partition under debian by giving fdisk the correct final cyl number. But I still had to use Mandrake fdisk to find this number! = Amateur Radio, when all else fails! http://www.qsl.net/wa2mze Debian Gnu Linux, Live Free or . __ Do You Yahoo!? Bid and sell for free at http://auctions.yahoo.com
Re: large hard disks (again)
The 2.2 kernels will properly detect 1024 cylinders. For other kernels, you can put fdisk in xpert mode and override its detection, or you can also specify the disk geometry on the kernel command line. Regards, Jeff On Wed, Sep 22, 1999 at 05:14:25AM -0700, Kenneth Scharf wrote: I have posted this before, how to fdisk a 8.4gb drive. Problem was that with slink fdisk saw only 1024 cy, 255h, 63s for a total of 8.4gb. If I booted a RedHat 6.0 or Mandrake 6.0 CD (also stormix) their partition utilities (fdisk, disk druid, etc) would report the correct number of cyl's (2100) to give a 17.2gb total disk size. After partitioning with Stormix I stopped the install, and booted my slink cd and then installed slink (skipping the partition step, which was already done). This worked. Now if I run fdisk, it still reports the wrong(?) number of cyl's but does show the correct partition sizes (p command). It also complains about different logical and physical starting and/or ending cyl numbers for the last two partitions on the disk. However if I go to expert mode (x) and change the number of cylinders (c) to 2100, the p command now gives the exact same printout as fdisk under Mandrake or Redhat fdisk. So...why does fdisk get the correct size info from the kernel under Mandrake or Redhat? Can it be a difference between the 2.0.36 kernel and the 2.2 series? Or is it a later version of fdisk itself (BTW cfdisk and sfdisk behave the same way...) I now think the difference is in the 2.2 kernel. Any ideas? I now know I can partition under debian by giving fdisk the correct final cyl number. But I still had to use Mandrake fdisk to find this number!
Re: large hard disks (again)
--- Jeff Noxon [EMAIL PROTECTED] wrote: The 2.2 kernels will properly detect 1024 cylinders. For other kernels, you can put fdisk in xpert mode and override its detection, or you can also specify the disk geometry on the kernel command line. Regards, Exactly what I suspected. Only how to figure out the geometry in the first place? Anyway I now have that information for the disk in question. As soon as potato is release it will be water under the bridge. = Amateur Radio, when all else fails! http://www.qsl.net/wa2mze Debian Gnu Linux, Live Free or . __ Do You Yahoo!? Bid and sell for free at http://auctions.yahoo.com
Re: large hard disks (again)
On Wed, Sep 22, 1999 at 10:15:37AM -0700, Kenneth Scharf wrote: Exactly what I suspected. Only how to figure out the geometry in the first place? Anyway I now have that information for the disk in question. As soon as potato is release it will be water under the bridge. Possible sources: - Label on the drive itself - Web site of drive manufacturer - BIOS (Award BIOS in particular has an auto-detect feature, which is not the same as setting the drive type as AUTO) - The partition table already on the drive, if the drive has already been partitioned by another OS. And others, I'm sure. Regards, Jeff
Re: large hard disks (again)
Well the right settings (as reported under the 2.2 kernel to fdisk) seem to be the number of cyl's that give the correct capacity with the head and sector numbers maxed out. (Ie 255 and 64). In my case for a 17.2 gb drive this was 2100. However this is NOT the physical geo, which IS reported by the bios and stamped on the drive. --- Jeff Noxon [EMAIL PROTECTED] wrote: On Wed, Sep 22, 1999 at 10:15:37AM -0700, Kenneth Scharf wrote: Exactly what I suspected. Only how to figure out the geometry in the first place? Anyway I now have that information for the disk in question. As soon as potato is release it will be water under the bridge. Possible sources: - Label on the drive itself - Web site of drive manufacturer - BIOS (Award BIOS in particular has an auto-detect feature, which is not the same as setting the drive type as AUTO) - The partition table already on the drive, if the drive has already been partitioned by another OS. And others, I'm sure. Regards, Jeff = Amateur Radio, when all else fails! http://www.qsl.net/wa2mze Debian Gnu Linux, Live Free or . __ Do You Yahoo!? Bid and sell for free at http://auctions.yahoo.com